x86/fpu: Fix i486 + no387 boot crash by only saving FPU registers on context switch...
authorSebastian Andrzej Siewior <bigeasy@linutronix.de>
Tue, 16 Oct 2018 20:25:25 +0000 (22:25 +0200)
committerIngo Molnar <mingo@kernel.org>
Wed, 17 Oct 2018 10:30:38 +0000 (12:30 +0200)
Booting an i486 with "no387 nofxsr" ends with with the following crash:

   math_emulate: 0060:c101987d
   Kernel panic - not syncing: Math emulation needed in kernel

on the first context switch in user land.

The reason is that copy_fpregs_to_fpstate() tries FNSAVE which does not work
as the FPU is turned off.

This bug was introduced in:

  f1c8cd0176078 ("x86/fpu: Change fpu->fpregs_active users to fpu->fpstate_active")

Add a check for X86_FEATURE_FPU before trying to save FPU registers (we
have such a check in switch_fpu_finish() already).

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Reviewed-by: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Fixes: f1c8cd0176078 ("x86/fpu: Change fpu->fpregs_active users to fpu->fpstate_active")
Link: http://lkml.kernel.org/r/20181016202525.29437-4-bigeasy@linutronix.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
arch/x86/include/asm/fpu/internal.h

index a38bf5a..69dcdf1 100644 (file)
@@ -528,7 +528,7 @@ static inline void fpregs_activate(struct fpu *fpu)
 static inline void
 switch_fpu_prepare(struct fpu *old_fpu, int cpu)
 {
-       if (old_fpu->initialized) {
+       if (static_cpu_has(X86_FEATURE_FPU) && old_fpu->initialized) {
                if (!copy_fpregs_to_fpstate(old_fpu))
                        old_fpu->last_cpu = -1;
                else